How Do Prognostic Models Factor into Treatment Decision Making for MDS?

A roundtable discussion moderated by Guillermo Garcia-Manero, MD, of the University of Texas MD Anderson Cancer Center, focused on the latest updates in lower-risk myelodysplastic syndromes (MDS). The panel included Jamie Koprivnikar, MD, of the Hackensack University Medical Center; Solly Chedid, MD, of Singing River Health System; and Thomas LeBlanc, MD, MA, of Duke Cancer Institute.

In the next roundtable segment, the panel discussed risk stratifying patients with MDS with the Revised International Prognostic Scoring System (IPSS-R) and the Molecular International Prognostic Scoring System (IPSS-M) and how that factors into treatment decision-making.

—

Dr. Garcia-Manero: How do you stratify patients? How do you use these prognostic models?

Dr. Koprivnikar: My personal practice has been to calculate both risk scores. The IPSS-M incorporates molecular data into our risk stratification and adds another level of risk stratification or gives us another category of risk compared to the IPSS-R. I am calculating both scores for my patients. I think it’s particularly important in transplant-eligible individuals. Data presented at the [American Society of Hematology Annual Meeting and Exposition] showed that the IPSS-M is a better prognostic model. It helps to more accurately identify patients who would benefit from earlier allogeneic [hematopoietic] stem cell transplant (HSCT). Particularly for that group of patients, it is important to incorporate that molecular data and use the IPSS-M.

However, I would also argue that when we’re deciding about a supportive approach with an agent like luspatercept, epoetin alfa, another [erythropoiesis stimulating agent (ESA)] versus more disease-modifying therapy with a [hypomethylating agent (HMA)], that even in a transplant-ineligible patient, risk stratifying that individual to the best of our abilities is going to help us bring the right treatment to the right patient.

Certainly, I am calculating both risk scores. You have those IPSS-R intermediate-risk patients, which sometimes could go either way. Again, I think that group of patients as we look at some of the molecular attributes of their disease and calculate their IPSS-M can really help us make a treatment decision going forward. Not to say that there aren’t other patient-specific factors like comorbidities, goals, quality of life that I integrate into my decision-making, but I think the IPSS-M gives us more accurate information, and that’s the direction the field is moving in.

Dr. Garcia-Manero: That’s a great answer. I do the same as you do. I calculate both. Sometimes, we see a patient with low-risk features like low percentage of blast, maybe the chromosomes are not that abnormal, maybe the cytopenia is not that severe. Then, you run your [next generation sequencing (NGS)] assay, and this person has a significant p53 mutation, and when you compare the IPSS-R with the IPSS-M, the risk goes up. From a clinical trial perspective, these are difficult because the IPSS-M has not yet been incorporated into clinical trial design. We are not saying you are higher risk by IPSS-M, low-risk by IPSS-R. We are not there yet.

Solly, how do you react to that in your practice?

Dr. Chedid: I feel like we’re diagnosing MDS in younger patients all the time. Getting [NGS] testing is substantially easier, and we’re doing it for everybody now with new onset diagnosis. We’re diagnosing these much, much earlier. Essentially, I’m referring these patients to our local transplant center. That’s the only curative modality that we have at this point. I think it’s reasonable, especially in young, healthy folks, that they at least have a discussion that you have some high-risk features. Even though an older model doesn’t appear that you would be in as bad a shape, potentially this disease could be substantially worse than it initially appears. It’s reasonable to at least have a discussion and have the person be in touch with a local transplant center so that that could be on their radar and you’re not scrambling to get them evaluated if and when things get worse.

Dr. LeBlanc: If I see a patient who has lower-risk disease by the IPSS-R and doesn’t have a high blast count or doesn’t need or benefit from disease-modifying therapy initially, those are the people we’ve always gone with supportive care approaches for. When we only had ESAs and transfusions, we’ve had patients who go for three, four, five plus years getting supportive care and have good quality of life, don’t progress to leukemia, and do well. You don’t want to sentence that person to four to six months of an HMA where you know they’re going to get worse. They might require transfusions, their counts drop, they may be in and out of the hospital with infections and complications, and then the response rate isn’t even that great to a single agent HMA.

If somebody’s lower-risk and it’s behaving that way clinically, I’m erring on the side of using all the supportive care therapies available now and reserving the more disease-modifying therapies for patients who have high blast counts and very high-risk disease, who benefit from transplant, who need to go to transplant where we’re trying to achieve disease control delay or reduce leukemic transformation and get them to hopefully a curative [allogeneic HSCT].

It’s the ones in between that are the hardest, where somebody doesn’t have a lot of blasts and their counts aren’t that bad and they’re feeling okay, but the IPSS-M tells me this is going to go badly. We don’t have the data that says you should give those folks HMAs, for example. It’s the classic story of prognostic models being useful to predict what happens to a population of patients, and they can be good at that, but that’s in 100 people. What’s going to happen to the person sitting across from you in the exam room? It never really tells you that. It just gives you a sense of which way to hedge your bets.

With the COMMANDS trial, it’s clear that using luspatercept as first-line therapy for most patients is our new standard of care. When we prognosticate, if somebody has lower-risk MDS and they’re expected to live five or eight plus years on median, we need every option we have. I want to use the best one first, which is luspatercept. Even if it works about the same for transfusion independence in ring-sideroblastic-negative [patients], the duration is longer. If you use it after [erythropoietin (EPO)], it doesn’t work nearly as well or for as long. I’ll use [luspatercept] first, and then I’m going to save the next bullet that we need against this disease, imetelstat, for the next line of therapy. That’s the way I’ve been thinking about it and approaching it.